Raveling (loss of aggregates) is one of the important asphalt pavement distresses. State DOTs like Florida DOT, Georgia DOT, and Alabama DOT use Open-graded Friction Course (OGFC) pavements; raveling is one of the most predominant pavement distresses on OGFC pavements. However, current practices for visual inspection of raveling severity levels are time-consuming, labor-intensive, and subjective.
NCHRP IDEA 20-30/IDEA 163 has successfully developed an automatic raveling classification algorithm using 3D pavement data, macro-texture analysis, and ML modeling. Different ML models have been critically evaluated; the most effective ML model has been identified, and an automatic raveling classification algorithm using ML, macro-texture analysis, and 3D pavement surface data has been developed.
The developed automatic raveling classification algorithm uses the 3D pavement surface data already collected by state Departments of Transportation (DOTs) for pavement evaluation of cracking and rutting, so no extra data collection effort is necessary. The output from the automatic raveling classification algorithm is the severity level (severe/3, medium/2, and low/1) based on the 3D pavement images collected at different image sizes (e.g., 5-meter or 8-meter intervals) as specified by the different 3D sensing systems used.
NCHRP 20-44(50) proposes to pilot the developed automatic raveling classification algorithm for improving the current visual inspection of raveling using 3D pavement data, and the following tasks are proposed:
- Prepare user manual and training materials, including briefing materials for senior management and flyers, brochures, and/or videos
- Perform a demonstration/a pilot study in a host agency using the tools and functions customized for state DOTs' needs.
- The demonstration will consist of aggregating data into sub-sections (e.g., 0.1-mile segments) and project level summaries. The two outputs will be used to identify short-term maintenance needs and longer-term rehabilitation requirements. Sub tasks include:
- Customizing the data processing tool to load and process a large quantity of 3D pavement image data and to store classification outcomes in a database with a standardized data format.
- Customizing the outcome quality checking tool for use by a host agency.
- Customizing the tool of aggregating the automatically generated image-based classified raveling outcomes to support state DOTs’ pavement management systems (PMS) and maintenance decisions.
- Conduct training and workshops on how to implement the developed automatic raveling classification algorithms and the developed tools and functions for incorporating the raveling classification outcomes into state DOTs’ PMS. This will include site visits and online trainings.